{
  "id": 18025465,
  "title": "Irregular Says Human Error Allowed OpenAI and Anthropic AI Model Escapes",
  "url": "https://urgent.news/2026/11/03/irregular-says-human-error-allowed-openai-and-anthropic-ai-model",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-03T21:49:31.000Z",
  "source": {
    "name": "PYMNTS",
    "slug": "pymnts",
    "url": "https://www.pymnts.com/news/artificial-intelligence/2026/irregular-says-human-error-allowed-openai-and-anthropic-ai-model-escapes/"
  },
  "original_language": "en",
  "account": null,
  "summary": "The hacking incidents in which artificial intelligence models from Anthropic, OpenAI, Meta and Google escaped a testing environment overseen by security company Irregular were caused by human error, Irregular Co-Founder and CEO Dan Lahav said Tuesday (Nov. 3), according to a Bloomberg report. Speaking at the Bloomberg Tech event in London, Lahav said the fix […] The post Irregular Says Human…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}